Yifan Zhu

Moi! I’m Yifan, currently an ELLIS PhD student at Aalto University and ELLIS Institute Finland, supervised by Prof. Samuel Kaski and advised by Dr. Sammie Katt. Through the ELLIS PhD & Postdoc Programme, I am also co-supervised by Dr. Mingfei Sun at the University of Manchester.

Driven by the vision of empowering human decision-making with AI, my research aims to align AI agents with users’ bounded rationality to support human agency. By developing computational and generative user models, I build AI agents that actively infer users’ characteristics at test-time, such as latent goals and cognitive bounds, and thereby provide adaptive assistance, ultimately contributing to more trustworthy, human-centric, and efficient collaborative AI systems.

This long journey goes back to my freshman year of studying computer science at Nanjing University of Aeronautics and Astronautics. Participating in an undergraduate research project on safe driving detection supervised by Prof. Bohan Li, I was introduced to Bayesian decision theory, and discovered my foundational interest in understanding human behaviours through probabilistic modelling. This spark was further strengthened during my exchange to the University of Edinburgh studying machine learning, and developed to human-centered AI during my thesis project on deep reinforcement learning with Prof. Kun Zhu. With the established technical foundation, I was inspired to study computational interaction and work on interactive AI with Prof. Antti Oulasvirta and Dr. Suyog Chandramouli at Aalto University, which ultimately shaped my research taste as well as the mission of my current PhD.